[PATCH] D108861: [X86] Remove X86LowerAMXType::getRowFromCol from X86LowerAMXType.cpp

[Vince Bridgers via Phabricator via llvm-commits]

vabridgers created this revision.
Herald added subscribers: foad, kerbowa, pengfei, hiraditya, tpr, mgorny, nhaehnle, jvesely, arsenm.
vabridgers requested review of this revision.
Herald added a project: LLVM.
Herald added a subscriber: llvm-commits.

Remove method X86LowerAMXType::getRowFromCol since it's not used, and
it's causing a warning.

[AMDGPU] Fix gcc 9.3.0 warning from -Wtype-limits

Fixes this warning when using gcc 9.3.0

lib/Target/AMDGPU/R600GenSubtargetInfo.inc:282:62: warning: comparison
of unsigned expression < 0 is always false [-Wtype-limits]
 if (Bits[R600::FeatureLocalMemorySize0] && LocalMemorySize <0)

  LocalMemorySize = 0;

[CMake] Disable -Wmisleading-indentation when using gcc

gcc doesn't handle -Wmisleading-indentation correctly, leading to
warnings when using gcc. See
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81334 for details. This
change disables -Wmisleading-indentation to avoid warnings when using
gcc, and leaves it enabled for other compilers. This can be re-enabled
when gcc behaves as expected.

[AMDGPU] Use gcc pragmas to disable -Wtype-limits warning

Use gcc pragmas to disable warning in generated filed. This is producing
a warning when using using gcc 9.3.0.

lib/Target/AMDGPU/R600GenSubtargetInfo.inc:282:62: warning: comparison
of unsigned expression < 0 is always false [-Wtype-limits]

  if (Bits[R600::FeatureLocalMemorySize0] && LocalMemorySize <0)
     LocalMemorySize = 0;


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D108861

Files:
  llvm/cmake/modules/HandleLLVMOptions.cmake
  llvm/lib/Target/AMDGPU/AMDGPUSubtarget.cpp
  llvm/lib/Target/AMDGPU/R600Subtarget.cpp
  llvm/lib/Target/X86/X86LowerAMXType.cpp


Index: llvm/lib/Target/X86/X86LowerAMXType.cpp
===================================================================
--- llvm/lib/Target/X86/X86LowerAMXType.cpp
+++ llvm/lib/Target/X86/X86LowerAMXType.cpp
@@ -175,25 +175,8 @@
   void combineLoadBitcast(LoadInst *LD, BitCastInst *Bitcast);
   void combineBitcastStore(BitCastInst *Bitcast, StoreInst *ST);
   bool transformBitcast(BitCastInst *Bitcast);
-  Value *getRowFromCol(Instruction *II, Value *V, unsigned Granularity);
 };
 
-Value *X86LowerAMXType::getRowFromCol(Instruction *II, Value *V,
-                                      unsigned Granularity) {
-  if (Col2Row.count(V))
-    return Col2Row[V];
-  IRBuilder<> Builder(&*II->getParent()->getFirstInsertionPt());
-  if (auto *I = dyn_cast<Instruction>(V)) {
-    BasicBlock::iterator Iter = I->getIterator();
-    ++Iter;
-    Builder.SetInsertPoint(&*Iter);
-  }
-  ConstantInt *Gran = Builder.getInt16(Granularity);
-  Value *RealRow = Builder.CreateUDiv(V, Gran);
-  Col2Row[V] = RealRow;
-  return RealRow;
-}
-
 // %src = load <256 x i32>, <256 x i32>* %addr, align 64
 // %2 = bitcast <256 x i32> %src to x86_amx
 // -->
Index: llvm/lib/Target/AMDGPU/R600Subtarget.cpp
===================================================================
--- llvm/lib/Target/AMDGPU/R600Subtarget.cpp
+++ llvm/lib/Target/AMDGPU/R600Subtarget.cpp
@@ -20,7 +20,10 @@
 
 #define GET_SUBTARGETINFO_TARGET_DESC
 #define GET_SUBTARGETINFO_CTOR
+#pragma GCC diagnostic push
+#pragma GCC diagnostic ignored "-Wtype-limits"
 #include "R600GenSubtargetInfo.inc"
+#pragma GCC diagnostic pop
 
 R600Subtarget::R600Subtarget(const Triple &TT, StringRef GPU, StringRef FS,
                              const TargetMachine &TM)
Index: llvm/lib/Target/AMDGPU/AMDGPUSubtarget.cpp
===================================================================
--- llvm/lib/Target/AMDGPU/AMDGPUSubtarget.cpp
+++ llvm/lib/Target/AMDGPU/AMDGPUSubtarget.cpp
@@ -37,7 +37,10 @@
 #define GET_SUBTARGETINFO_TARGET_DESC
 #define GET_SUBTARGETINFO_CTOR
 #define AMDGPUSubtarget GCNSubtarget
+#pragma GCC diagnostic push
+#pragma GCC diagnostic ignored "-Wtype-limits"
 #include "AMDGPUGenSubtargetInfo.inc"
+#pragma GCC diagnostic pop
 #undef AMDGPUSubtarget
 
 static cl::opt<bool> DisablePowerSched(
Index: llvm/cmake/modules/HandleLLVMOptions.cmake
===================================================================
--- llvm/cmake/modules/HandleLLVMOptions.cmake
+++ llvm/cmake/modules/HandleLLVMOptions.cmake
@@ -789,7 +789,11 @@
   add_flag_if_supported("-Wstring-conversion" STRING_CONVERSION_FLAG)
 
   # Prevent bugs that can happen with llvm's brace style.
-  add_flag_if_supported("-Wmisleading-indentation" MISLEADING_INDENTATION_FLAG)
+  if (CMAKE_COMPILER_IS_GNUCXX)
+    add_flag_if_supported("-Wno-misleading-indentation" MISLEADING_INDENTATION_FLAG)
+  else()
+    add_flag_if_supported("-Wmisleading-indentation" MISLEADING_INDENTATION_FLAG)
+  endif()
 endif (LLVM_ENABLE_WARNINGS AND (LLVM_COMPILER_IS_GCC_COMPATIBLE OR CLANG_CL))
 
 if (LLVM_COMPILER_IS_GCC_COMPATIBLE AND NOT LLVM_ENABLE_WARNINGS)


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D108861.369243.patch
Type: text/x-patch
Size: 3083 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20210828/9a4540e2/attachment.bin>